AMBI and M-AMBI for use in National Coastal Condition Surveys and mAMBI - Pelletier.pdf · • AMBI...
Transcript of AMBI and M-AMBI for use in National Coastal Condition Surveys and mAMBI - Pelletier.pdf · • AMBI...
Marguerite (Peg) PelletierEPA ORD National Health and Environmental Effects Research Laboratory
Atlantic Ecology Division
AMBI and M-AMBI for use in National Coastal Condition Surveys
West Coast Salinity-adjusted expected number of species
Existing Coastal Indices
• Benthic Indices developed and calibrated separately for each Region• Concerns about cross-region comparability
Existing Coastal Indices
BI = 0 * % EG I + 1.5 * % EG II + 3 * % EG III + 4.5 * EG IV + 6 * EG V
Range = 0 (unimpacted) to 6 (heavily impacted) or 7 (azoic)
AMBI – Initial Development
• Previous Case Studies in FL, SoCal, Chesapeake Bay, Northwest
• 3 Day workshop (Sept 2011)• NCCA species categorized by EG group• Workshop EG list augmented with existing European EG
list• 3 regional Datasets assembled – compared local index to
AMBI• Published results in 2015
AMBI – Adaptation to US estuaries
R = 0.736, p<0.0001 R = 0.525, p<0.0001
R = 0.437, p<0.0001
AMBI – Adaptation to US estuaries
AMBI – Adaptation to US estuaries
• Strong salinity bias seen in unimpacted station in the southeast and mid-Atlantic (SoCal is primarily high salinity sites so a salinity bias would not be expected)
• Used Factor Analysis to combine AMBI, diversity (H’) and species richness into a new index
• Classified by habitat (salinity and location (coastal)• Good and Bad endpoints derived for each metric• Range = 0 (Bad) to 1 (High)
M-AMBI – Initial Development
M-AMBI – Adaptation to US estuaries
• Venice salinity classification to identify habitats
• Examined West Coast data – used larger grab (0.1 m2 grab vs 0.04 m2 grab or equivalent) and sieve (1.0 mm vs. 0.5 mm)
M-AMBI – Adaptation to US estuaries
• West Coast gear differences• Sieve size differences did not appear to be significant based on subset of stations
using both sieves
• Grab size impacted the total number of species
• Derived Bad/High endpoints for factor analysis calculation based on habitat• Tidal freshwater, Oligohaline, Mesohaline, Hyperhaline – few of these sites on West
coast, so calculated for entire US• Polyhaline and Euhaline – calculated thresholds separately for West only and Rest of
US
• Used raw abundance (rather than ln(abundance) as in Gillett paper) due to dampening of benthic response to chemical contamination
M-AMBI – Adaptation to US estuaries
Pearson & Rosenberg 1978
M-AMBI – Adaptation to US estuaries
• Using 2000-2006 NCA data, explored use of metrics other than S and H’ (e.g., dominance % oligochaetes)
• Calculating M-AMBI for 3 validation datasets• Compare to local indices• Look at calibration accuracy vs. apriori Good/Bad sites• Look to see if salinity correlation has been reduced or eliminated
Summary
• AMBI is an abundance-weighted tolerance index analogous to the Hilsenhoff Index (conceptually)
• M-AMBI is multivariate AMBI• Accounts for naturally structuring parameters (e.g., salinity)• Improves index performance by adding additional metrics
• For the Great Lakes we would like to have an index that is conceptually compatible with the estuarine approach, if possible